All tags
Topic: "low-latency"
not much happened today
gemini-3.1-flash voxtral-tts cohere-transcribe gpt-5.4-mini gpt-5.4-nano glm-5-turbo reka-edge reka-flash-3 google-deepmind mistral-ai cohere openai zai reka-ai voice vision function-calling context-windows multimodality text-to-speech low-latency human-preference automatic-speech-recognition model-benchmarking cost-efficiency hallucination-detection multi-agent-systems open-source git-worktrees logan_kilpatrick sundar_pichai guillaume_lample aidan_gomez jay_alammar giffmana andrew_curran
Google launched Gemini 3.1 Flash Live, a realtime voice and vision agent model with 2x longer conversation memory, supporting 70 languages and 128k context. Mistral AI released Voxtral TTS, a low-latency, open-weight text-to-speech model supporting 9 languages and competitive with ElevenLabs. Cohere introduced Cohere Transcribe, an audio model with 14-language support and top English ASR leaderboard performance at 5.42 WER. OpenAI released smaller multimodal variants GPT-5.4 mini and GPT-5.4 nano with 400k context, noted for cost-competitiveness but high verbosity and hallucination rates. Other releases include GLM-5-Turbo by Zai, Reka Edge and Flash 3 on OpenRouter, and new multi-agent UX tooling Cline Kanban for orchestrating CLI coding agents.
Nvidia buys (most of) Groq for $20B cash; largest execuhire ever
gemini fsd-v14 nvidia groq openai tesla epoch-ai gemini benchmarking inference model-evaluation ai-integration agent-patterns real-time-processing low-latency developer-experience healthcare business-workflows consumer-ai jensen_huang xeophon js_denain jim_fan
Groq leadership team is joining Nvidia under a "non-exclusive licensing agreement" in a deal valued at $20 billion cash, marking a major acquisition in AI chip space though Nvidia states it is not acquiring Groq as a company. Jensen Huang plans to integrate Groq's low-latency processors into the NVIDIA AI factory architecture to enhance AI inference and real-time workloads. Twitter highlights include Gemini used as a consumer utility for calorie tracking, OpenAI discussing the "deployment gap" focusing on model usage in healthcare and business, and Tesla's FSD v14 described as a "Physical Turing Test" for consumer AI. Benchmarking challenges are noted by Epoch AI emphasizing provider variance and integration issues affecting model quality measurement. Discussions on coding agents and developer experience convergence continue in the AI community.